The tree nearest on average to a given set of trees

We formulate the problem of constructing a tree which is the nearest on average to a given set of trees. The notion of “nearest” is formulated based on a conception of events such that counting their number makes it possible to distinguish each of the given trees from the desired one. These events are called divergence, duplication, loss, and transfer; other lists of events can also be considered. We propose an algorithm that solves this problem in cubic time with respect to the input data size. We prove correctness of the algorithm and a cubic estimate for its complexity.